Table of Contents
In a joint trial, Telstra and Silicon Quantum Computing demonstrated that a quantum system can forecast network issues faster than a deep learning model
In sum – what to know:
Quantum meets machine learning — Telstra and Silicon Quantum Computing jointly tested a quantum machine learning system for network automation.
Efficiency edge – The system, while running on less powerful hardware than AI models, performed on par with Telstra’s deep learning model.
Early, but promising – QML might not be ready for primetime yet, but the test marks a meaningful step in adoption of quantum computing in telecommunication.
Engineers at Sydney-based telecom company, Telstra, have completed a 12-month long trial with Silicon Quantum Computing’s (SQC) quantum machine learning (QML) system. The results are in — and they are promising.
As the telecom industry pushes towards next-generation connectivity like 6G, demand for AI-driven service optimization, predictive analytics, and stronger network security prompt adoption of progressive technologies like quantum computing, that go beyond the limits of classical computing.
For the past few years, telco companies have been testing QML in their labs as a potential technology to use in telecommunication settings. QML — a mishmash of quantum computing and machine learning — applies quantum computing on machine learning applications to enhance model accuracy and solution quality through a quantum boost. QML is especially well-suited to make sense of noisy data, and has proven to deliver superior accuracy in early experiments.
Aleena Taufiq, Principal Engineer at Verizon, at the Ai4 2024 summit called QML the “coolest new tech related to AI” and said: We can expect to see some small scale production applications within the next three years.
The Telstra-SQC joint initiative aimed to take quantum computing out of the lab and explore its real-world potential in predictive network analytics. The test used SQC’s quantum-based machine learning system, Watermelon — a quantum feature generator — that is designed to identify hidden insights in data.
According to Telstra, Watermelon detected pattern changes in Telstra’s network traffic and predicted performance issues with high accuracy. The system’s output even matched that of Telstra’s own newly developed deep learning model. Notably, the system trained much faster, and required less powerful hardware compared to conventional AI models.
“The quantum-enhanced model matched the deep learning system on key tasks during the trials. What stood out: the quantum reservoir trained in just days – versus weeks for the deep learning model – showing early promise in speed and efficiency,” Kim Krogh Anderson, Product and Technology Group Executive, wrote in a LinkedIn post.
AI/ML models are already widely leveraged by telecom operators worldwide. Providers like Vodafone, Verizon, AT&T and Comcast rely on AI and machine learning models to anticipate equipment failures, downtimes and disruptions, and react to network problems proactively.
Telstra too currently uses AI and machine learning models to monitor network metrics like latency, bandwidth, and jitters. Integration of these tools allows it to anticipate and address problems before they hit the users.
The quantum-enabled model potentially takes it to the next level. The test proves that features generated by Watermelon could be practically applied to analyze network metrics and forecast performance issues, essentially powering network automation.
“Telstra is not a quantum company or an AI company – but to lead in connectivity and succeed with our Connected Future 30 strategy, we need to be a leader in both,” Anderson wrote in his post.
As breakthroughs like this keep surfacing, hopes of commercial adoption is growing with it. And an emerging class of quantum test and measurement tools is accelerating innovation, bringing it closer to reality. However, it is unclear whether Telstra’s model is production-ready today, or if it needs more refinement before it can be deployed at scale. Regardless, the project stands out as one of Australia’s earliest attempts to move quantum computing out of its experiment bed and into the real-world digital infrastructure for operational use.
“This trial shows how quantum capabilities could complement our existing systems and technology to deliver faster insights and better outcomes for our customers,” Shailin Sehgal, Group Executive of Global Networks and Technology at Telstra said.
Telstra had not yet responded to a request for comment at the time of writing.
